Robustness in SSD portfolio efficiency testing

This paper deals with portfolio efficiency testing with respect to secondorder stochastic dominance (SSD) criterion. These tests classify a given portfolio as efficient or inefficient and give some additional information for possible improvements. Unfortunately, these tests usually assumes discrete probability distribution of returns with equiprobable scenarios. Moreover, they can be very sensitive to the probability distribution. Therefore, [1] derived a SSD portfolio efficiency test for a general discrete distribution and its robust version allowing for changes in probabilities of the scenarios. However, these formulations are not suitable for numerical computations. Therefore, in this paper, the numerically tractable reformulations of these SSD efficiency tests are derived.